Understanding Housing Market Volatility
利用动态因子模型分解23个主要住房市场的租金价格比，发现大部分波动是地方性的，但自1999年以来全国性因素变得更重要；地方波动主要源于风险溢价变化，而非增长；总体而言，增长和利率因素只能解释不到一半的波动，其余来自风险溢价和定价错误，后者与货币幻觉有关。
The Campbell–Shiller present value formula implies a factor structure for the price–rent ratio of housing market. Using a dynamic factor model, we decompose the price–rent ratios of 23 major housing markets into a national factor and independent local factors, and we link these factors to the economic fundamentals of the housing markets. We find that a large fraction of housing market volatility is local and that the national factor has become more important than local factors in driving housing market volatility since 1999, consistent with the findings in Del Negro and Otrok (2007). The local volatilities mostly are due to time variations of idiosyncratic housing market risk premia, not local growth. At the aggregate level, the growth and interest rate factors jointly account for less than half of the total variation in the price–rent ratio. The rest is due to the aggregate housing market risk premium and a pricing error. We find evidence that the pricing error is related to money illusion, especially at the onset of the recent housing market bubble. The rapid rise in housing prices prior to the 2008 financial crisis was accompanied by both a large increase in the pricing error and a large decrease in the housing market risk premium.